|Publication number||US4533964 A|
|Application number||US 06/432,384|
|Publication date||Aug 6, 1985|
|Filing date||Sep 30, 1982|
|Priority date||Oct 15, 1981|
|Also published as||DE3237789A1|
|Publication number||06432384, 432384, US 4533964 A, US 4533964A, US-A-4533964, US4533964 A, US4533964A|
|Inventors||Isao Owaki, Susumu Saito, Masami Yamazaki|
|Original Assignee||Victor Company Of Japan, Limited|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (10), Non-Patent Citations (2), Referenced by (9), Classifications (8), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to a digital recording and playback system in which digital signals of non-return-to-zero inverted format are recorded at high density using a plurality of stationary magnetic heads.
In digital recording systems known in the art, audio signals are translated to PCM signals and then organized into a sequence of blocks each comprising digital words of data and synchronization and digital words of error detection and correction. The binary digits of the organized signal are of non-return-to-zero (NRZ) format. However, due to the fact that the clock component of the NRZ signal tends to vary with the contents of the data word, the clock information would be lost if a sequence of zero's continues. A transformation technique is currently employed to transform the original non-return-to-zero (NRZ) signal into a form which contains a sufficient amount of clock information to make it highly likely that binary digits are reproduced in correct clock timing.
On the other hand, the use of conventional 3.81-mm width tapes and conventional stationary heads having a minimum gap of 0.3 micrometers imposes limitation on recording density although high tape-speed operations would solve the density problem.
Known transformation techniques include modified frequency modulation (MFM), 3-position modulation (3PM), and group-coded recording (GCR). Since MFM or 3PM transformed, self-clocked signals have a wide frequency spectrum containing a substantial amount of energy the very low frequency range near DC and in the high frequency range, conventional recording systems having a typical tape speed of 7.1 cm/sec suffer from cross-talk due to leakage flux between adjacent recording heads, with the result that the reproduced waveform is contaminated with noise and the error rate increases considerably in the high frequency range. Moreover, these problems are compounded by the fact that due to the differential characteristic the magnetic head produces a ternary signal rather than a binary signal, making it difficult to achieve high density recording.
The present invention contemplates the elimination of the above-mentioned problems by transforming the digital signal into a non-return-to-zero inverted (NRZI) signal, distributing the signal to a plurality of parallel stationary heads and detecting the sync word for clock regeneration and permits high density recording along a first set of multiple tracks of a magnetic tape when driven toward one end thereof and along a second set of multiple tracks interleaved with the first tracks when the tape is driven toward the other end thereof.
According to the invention, the digital recording and playback system comprises means for converting an analog signal to a digital signal comprising a data word and a synchronization word, means for distributing the digital signal to a plurality of circuit paths, means for transforming the digital signal on each circuit path into a non-return-to-zero inverted (NRZI) signal, a plurality of stationary parallel electromagnetic transducers respectively associated with the circuit paths and spaced apart at least by the width of a track for recording the distributed NRZI signals along a plurality of longitudinally extending first tracks of a magnetic tape when the tape is driven toward one end thereof and along a plurality of longitudinally extending second tracks interleaved with the first tracks when the tape is driven toward the other end thereof, each of the transducers having a differential characteristic with which the recorded signal is detected as a waveform having three significant levels when the system operates in a playback mode, means coupled to the transducers for converting the waveform to a binary signal to recover the original digital signal, means for detecting the synchronization word from the recovered digital signal to generate a timing signal indicative of the position of each binary digit of the data word with respect to the detected synchronization word, means responsive to the timing control signal for aligning the timing of each binary digit derived from each transducer with those derived from the other transducers, means for arranging the time-aligned data words into a series of successive data words, and means for converting the series of data words to an analog signal.
The invention permits the use of conventional tape recording mechanisms to attain a high recording density that allows the signal to be recorded at a rate of as high as 2 Megabits per second.
The present invention will be described in further detail with reference to the accompanying drawings, in which:
FIG. 1 is an illustration of a digital recording system of the invention;
FIG. 2 is a schematic illustration of a recording pattern;
FIG. 3 is an illustration of the waveform of a non-return-to-zero inverted signal;
FIG. 4 is an illustration of the frequency spectrum of the NRZI signal;
FIG. 5 is an illustration of a track pattern;
FIG. 6 is an illustration of a playback system of the invention;
FIG. 7 is an illustration of details of a level detector and a data readout circuit of FIG. 6; and
FIGS. 8 and 9 are illustrations of timing diagrams associated with the circuit of FIG. 7.
Referring now to FIG. 1, there is shown a preferred embodiment of the digital recording system of the present invention. Analog stereophonic signals are applied to right and left channels of the system. The right channel signal R is applied to a low-pass filter 1R where the bandwidth is limited to one-half the sampling frequency at which the audio signal is sampled in a sample-hold circuit 2R in a known manner; the sampled value being converted to a corresponding digital value in an analog-digital converter 3R to generate a pulse code modulated (PCM) signal, or a sequence of data words. Similarly, the left channel signal L is processed through an identical circuit comprising a low-pass filter 1L, a sample-hold circuit 2L, and an analog-digital converter 3L.
The right and left channel PCM data words are applied to a distribution and interleaving circuit 4 in which the data words of each audio channel are distributed to a number of circuit paths that correspond to parallel tracks (which, in this embodiment, is eight) on a recording tape and then interleaved with the distributed data words of the other audio channel. A parity word generator 5 is coupled to the interleaving circuit 4 to generate P and Q parity words according to the following Equations:
P=D1 W1 ⊕D2 W2 ⊕ . . . ⊕D14 W14
Q=D15 W1 '⊕D16 W2 '⊕. . . ⊕D28 W14 '⊕D29 P'
where, D1 to D14 represent delays A and D15 to D29 represent delays B, W1 to W14 and W1 ' to W14 ' represent data words, and the symbol ⊕ represents modulo-2 summation. The data words are sufficiently spaced apart on record tracks to make it highly unlikely that more than one data word will be affected by the same dropout defect and the redundant parity words are recorded on a track adjacent to an edge of the tape which is more subject to noise and fluctuation than the inner tracks, one example being shown in FIG. 2. The interleaved signals are applied to a frame sync and CRC generator 6. The data and parity words of each track are assembled with a frame synchronization word (00110110) to form a block, or frame to which is appended a cyclic redundancy check word derived from the data words of each frame in a well known manner by a generator polynomial.
The assembled digital words for each track, which are of non-return-to-zero format, are supplied to a respective precoder 7, for simplicity only one precoder being shown in detail. The precoder 71 comprises a one-bit delay element 8, an inverter 9 and an adder 10 having a first input connected to receive the NRZ input signal directly from the generator 6 and a second input connected to receive a one-bit delayed, inverted NRZ signal passing through the delay element 8 and inverter 9.
Each of the precoders 7 has a characteristic known as "partial response" which serves to convert the input NRZ signal to a non-return-to-zero inverted (NRZI) signal. More specifically, the output of the precoder 71 changes its pulse level from "0" to "1" or vice versa depending on the previous pulse level in response to the occurrence of each "1" bit of input NRZ signal and retains the pulse level when the input signal is "0", as shown in FIG. 3. The outputs of the precoders 71 to 78, after having been amplified by amplifiers 11, are supplied to respective transducer heads 121 to 128 of the in-line gap type which are arranged in parallel across the width of a magnetic tape 13 and spaced apart at least by the width of a track to record the NRZI signals along tracks T1 to T8 parallel to the length of the tape when driven toward one end thereof and along tracks T1' to T8' when the tape is driven toward the other end thereof as illustrated in FIG. 5. Each track has a width of typically 150 micrometers and is spaced from adjacent tracks by a guard band of typically 40 micrometers. A typical value of recording density is 46.3 kilobits per inch. Since the recorded NRZI signal has no DC component, the interference between adjacent magnetic heads is minimized.
FIG. 6 is an illustration of the playback section of the system. The magnetic heads 121 to 128 are used in reproducing the recorded signals. These magnetic heads have a differential characteristic with which the signal is differentiated upon reproduction. As a result, the output of each transducer head is a ternary signal rather than binary as shown at a in FIG. 8. The signals detected by the heads 121 to 128 are fed to a circuit 14 for amplification and equalization and thence to level detectors 151 to 158, respectively. Each level detector is arranged to detect the levels of the ternary signals and provide a pulse in response thereto. The outputs of the level detectors 151 to 158 are fed to frame and bit synchronizers 161 to 168 to recover the original binary signals and derive control timing pulses for use in a time base corrector 17.
Details of the level detector and frame and bit synchronizer are shown in FIG. 7. The level detector 15 comprises a pair of comparators 22 and 23 and an OR gate 24 having its inputs coupled respectively to the outputs of the comparators 22 and 23. The noninverting and inverting inputs of the comparators 22 and 23 are coupled together to an input terminal 21 to which an output signal a (FIG. 8) from the circuit 14 is applied. The frequency spectrum of the signal a of FIG. 8 is shown at FIG. 4. The ternary signal a compared with a reference voltage E1 in the comparator 22 to provide pulses b when the input level is higher than E1 and compared with a reference voltage E2 in the comparator 23 to provide pulses c when the input level is lower than E2. Pulses b and c are combined in the OR gate 24 to generate a train of pulses d. The pulses d are applied to the data terminals of D-type flip-flops 25 and 26. As can be seen, the output pulses d more quickly rise to a high voltage level when the ternary waveform a changes from "0" to "1" and delay before rising to the high level when changing from a given "1" binary level to the next "1" binary level.
The D-type flip-flop 25 has its Q output terminal connected to the data terminal of a D-type flip-flop 28 and the flip-flops 25 and 28 have their clock terminals connected together to a clock pulse source 27 to receive clock pulses at a rate which is 16 times higher than the data bit rate. The Q output of the flip-flop 25 and the complementary output of the flip-flop 28 are connected to an AND gate 29 to generate narrow pulses e in response to the leading edge transition of pulses d as shown in FIG. 8.
A divide-by-16 counter 31 is provided having a clock input CK, data input terminals D1 to D4, a load command input terminal LD, a carryover output terminal CA and binary output terminals Q1 to Q4. The data input terminal D1 is a least significant bit position, which together with the next higher bit terminal D2 are coupled to a source that provides a binary "0" potential, and the data input terminals D3 and D4 are respectively coupled to the complementary and true outputs of the flip-flop 26. The counter 31 has its clock input CK coupled to the clock source 27 to increment its count to cause an output pulse to occur on terminal CA at a rate corresponding to the data bit rate, this output pulse being applied to the enable terminal EN of a counter 34. The binary output terminals Q1 to Q4 are coupled to the corresponding input terminals of a decoder 32 having output terminals P1 and P2. The output of AND gate 29 and the terminal P2 of the decoder 32 are coupled by an AND gate 30 to the load terminal LD of the counter 31 to set the data input terminal D3 or D4 to decimal 4 or 8 respectively. The divide-by-16 counter 31 is set to a decimal "4" in response to a pulse e1 when the previous binary level of the waveform a is ".0." and initialized to a decimal "8" in response to a pulse e2 if the previous level is "1". This is to compensate for the difference in rise time as mentioned above so that it ensures that the D-type flip-flop 26 read its data during the period of each data bit.
Decoder 32 provides a trigger pulse f on terminal P1 to the clock input of the D-type flip-flop 26 when the count value reaches a decimal "1 ", for example. Count operation starts in counter 31 in response to pulse e1 so that the count begins with decimal 4 and terminates at decimal 11 when waveform a is previously ".0.", while it starts with decimal 8 in response to pulse e2 and terminates at decimal 11 when wave a is previously "1". Thus, a trigger pulse f1 occurs at a time delayed by an interval corresponding to a count of decimal 7 and a trigger pulse f2 occurs at a time delayed by an interval corresponding to a count of decimal 3. In the absence of initializing pulse e, count operation continues, incrementing the count value to decimal 15 and then to decimal .0. to increase the count again and repeating this process indefinitely. Thus, in the absence of pulses d, trigger pulses fa , fb, and fc are generated by decoder 32 each time decimal 11 is reached in counter 31 at times delayed by an equal interval corresponding to a count of decimal 16 with respect to their respective preceding trigger pulses f. It is appreciated that trigger pulses f1, f2 occur at times delayed an variable period with respect to the leading edge of pulses d1 and d2, respectively, which period is substantially equal to one half the constant intervals between pulses fa and fb and between f2 and fc when Q output of flip-flop 26 is low and which period is smaller than one half these intervals when the Q output is high. As is seen from FIG. 8, the intervals between trigger pulses fa and fb and between f2 and fc are equal to the period that determines the transmission bit rate of the input data a. The Q output of flip-flop 26 changes to the binary state of the pulse d in response to the pulse f to provide pulses g to an output terminal 33 and to the data terminal D4 of counter 31. The decoder 32 provides a high level signal from terminal P2 to the AND gate 30 during the interval when the count value of counter 31 is within the range of from a decimal "2" to a decimal "10", for example, to eliminate errors which would otherwise result from the level fluctuation of the input signal.
The counter 34, when enabled, starts counting clock pulses from source 27 at correct timing to generate a timing control binary output on terminals 36 indicating the location of each data bit with respect to the beginning of the data block to which the data bit belongs. The contents of the counter 34 are cleared in a manner described hereinbelow.
Since the ternary waveform a does not lends itself to regenerating clock pulses because of the inherent nature of the NRZI signal. In order to establish a valid timing for timing control purposes, the system repeatedly detects the occurrences of frame sync words by including a sync detector 40, a timer 41, a monostable multivibrator 42 and a logic circuit formed by AND gates 43, 44 and an OR gate 45. The sync detector 40 essentially comprises a shift register and a logic gate coupled to the internal stages of the register. The shift register is arranged to accept the output pulse g from flip-flop 26 and shift its contents in response to the output on terminal CA of the divide-by-16 counter 31 to detect a sync word when it corresponds to the bit pattern defined by the logic circuit. From the logic circuit is generated a pulse h as shown in FIG. 9.
Timer 41 is arranged to count the clock pulse to provide a timing pulse at frame intervals to a monostable multivibrator 42 to cause it to produce a window gate pulse i as illustrated in FIG. 9 for application to an AND gate 43 to which the pulse h is also applied. By application of the window gate pulses, only those of the detected pulses h which are valid for clock timing are passed through OR gate 45 to the clear terminals of the counter 34 and the timer 41.
A manually operated transfer switch SW is connected to an input of AND gate 44 to which the outputs of the sync detector 40 and timer 41 are also applied. The switch SW is normally connected to a "0" bit position and switched to a "1" bit position when the system is operated. With the switch SW being transferred to the "1" position, the AND gate 44 is enabled to pass an output pulse upon coincidence between a pulse h and a timing pulse from the timer 41, the coincidence output being used to clear the timer 41 during starting periods. Once the system is synchronized, a logical "0" is applied to the AND gate 44 by switching the switch SW to the "0" position to allow the timer 41 to be cleared in response to the output of AND gate 43.
In this way, the counter 34 is cleared in valid timing with respect to the input signal and the timing control signal on terminal 36 is indicative of the correct location of each data bit. Clock timing is thus achieved with a high degree of stability not attainable with conventional systems.
The signals on terminals 33 and 36 are applied to the time base corrector 17. The time base corrector 17 essentially comprises one or more memories in which the data bits from the frame and bit synchronizers 161 to 168 are stored and read out the memories in accordance with the timing control signals on terminals 36 to compensate for any difference in timing which might occur between signals from tracks as a result of a tape being skewed with respect to the aligned heads.
The outputs of the time base corrector 17 are then applied to a de-interleaving and error correction circuit 18 where data bits in errors, if any, are replaced by correct ones using CRC, P and Q parity words and the parallel data streams are de-interleaved into the original data. The de-interleaved data words are then fed to digital-analog converters 19R, 19L, and low-pass filters 20R and 20L to recover the original right and left channel signals.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4146099 *||Aug 11, 1977||Mar 27, 1979||Christopher Scientific Company||Signal recording method and apparatus|
|US4211997 *||Nov 3, 1978||Jul 8, 1980||Ampex Corporation||Method and apparatus employing an improved format for recording and reproducing digital audio|
|US4238852 *||Apr 16, 1979||Dec 9, 1980||Sony Corporation||Error correcting system|
|US4274115 *||May 9, 1979||Jun 16, 1981||Matsushita Electric Industrial Co., Ltd.||Apparatus for detecting recording and reproducing state|
|US4323931 *||Jul 3, 1980||Apr 6, 1982||Sperry Corporation||Method and apparatus for encoding and recovering binary digital data|
|US4376268 *||Oct 23, 1980||Mar 8, 1983||Sony Corporation||Phase locked loop having automatic-input tuning and phase control circuits|
|DE2827826A1 *||Jun 24, 1978||Jan 3, 1980||Bosch Gmbh Robert||Verfahren zur aufzeichnung von impulsen|
|EP0048933A1 *||Sep 22, 1981||Apr 7, 1982||Hitachi, Ltd.||Circuit for correcting error in digital information signal|
|FR2358786A1 *||Title not available|
|GB2056206A *||Title not available|
|1||*||D. Thomson, Digitale Audiotechnik, 1983, Franzis Verlag GmbH, Munich, pp. 80/81.|
|2||D. Thomson, Digitale Audiotechnik, 1983, Franzis-Verlag GmbH, Munich, pp. 80/81.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US4581759 *||Jun 26, 1984||Apr 8, 1986||Nippon Gakki Seizo Kabushiki Kaisha||Signal delaying device|
|US5109303 *||Dec 4, 1989||Apr 28, 1992||Zymed||Apparatus for accurate time alignment of signal data derived from multiple channels|
|US5257107 *||Dec 28, 1990||Oct 26, 1993||Samsung Electronics Co., Ltd.||Digital signal recording system for recording digital signals on video tape|
|US5313342 *||Nov 20, 1992||May 17, 1994||Sony Corporation||Analog/digital compatible reproducing system|
|US5357378 *||Oct 9, 1992||Oct 18, 1994||Sony Corporation||Data reproducing apparatus with means for differentiating a read head output signal and means for three value detection of the result differentiated|
|US5703645 *||Aug 30, 1990||Dec 30, 1997||Hitachi, Ltd.||Video signal transmitting apparatus|
|US7711036 *||Sep 8, 2005||May 4, 2010||Oki Semiconductor Co., Ltd.||Synchronous acquisition circuit and a synchronous acquisition method of a spread spectrum code|
|US7911730 *||May 7, 2008||Mar 22, 2011||Oracle America, Inc.||Magnetic head and tape drive for reading multiple data formats|
|US20060072653 *||Sep 8, 2005||Apr 6, 2006||Takamitsu Hafuka||Synchronous acquisition circuit and a synchronous acquisition method of a spread spectrum code|
|U.S. Classification||360/22, G9B/20.049, 360/26, 360/51|
|International Classification||G11B20/18, G11B20/12|
|Sep 30, 1982||AS||Assignment|
Owner name: VICTOR COMPANY OF JAPAN, LIMITED 3-12 MORIYA-CHO,
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:OWAKI, ISAO;SAITO, SUSUMU;YAMAZAKI, MASAMI;REEL/FRAME:004056/0101
Effective date: 19820927
|Mar 7, 1989||REMI||Maintenance fee reminder mailed|
|Mar 9, 1989||REMI||Maintenance fee reminder mailed|
|Aug 6, 1989||LAPS||Lapse for failure to pay maintenance fees|
|Oct 24, 1989||FP||Expired due to failure to pay maintenance fee|
Effective date: 19890806